Beam Help
Get help now

How-to · Zoho CRM

How to retrieve ZIA predictions in Zoho CRM

Get AI-powered predictive insights for your records.

Retrieving Zia predictions in Zoho CRM lets you surface AI-generated scores and forecasts — such as churn likelihood or deal win probability — for any given record programmatically or through the platform's built-in AI layer.


Why this matters


When your team needs to act on AI-driven insights at scale, manually checking each record inside the CRM UI is impractical. Pulling Zia predictions via a tool or API call means you can embed those scores into dashboards, trigger automations, or feed them into external reporting pipelines. Zia's prediction capabilities cover a range of signals, from purchase patterns to communication touchpoints, making programmatic access especially valuable for revenue and retention workflows. [3]


Step-by-step


Step 1. Confirm that Zia's Prediction feature is enabled for your Zoho CRM edition. Zia capabilities — including Prediction — are available across specific editions, so verify your plan covers this feature set before proceeding. [1][2]


Step 2. Identify the module and the record ID you want predictions for. Zia predictions are scoped to individual records (for example, a specific Lead, Contact, or Deal), so you will need both the module name (m) and the record identifier (rid) ready before making any call. [5]


Step 3. Invoke the getziapredictions tool, passing the module and record ID as parameters. The tool signature is:


get_zia_predictions(m, rid)

Where m is the module name and rid is the unique record ID. This tool is registered under the crm service and is categorised within Zoho CRM's Zia (AI) capability set. [5]


Step 4. Parse the response returned by getziapredictions. The output will contain Zia's scored predictions for that record — for instance, a churn score derived from purchase patterns, communication history, and any usage data you have mapped into Zoho CRM. [3]


Step 5. If you want to improve prediction accuracy over time, consider feeding usage data into Zia via API, URL, or a supported file format stored in a custom module. You can map events and usage metrics so that Zia has richer signals to work with when generating future predictions. [3]


Step 6. For forecast-level anomaly detection (rather than record-level predictions), note that Zia also monitors overall sales performance during a forecasting period and will automatically flag deviations to sales leaders — this is separate from the per-record getziapredictions call but is part of the same Zia intelligence layer. [4]


Common pitfalls


  • Missing module or record ID. The getziapredictions tool requires both m and rid. Passing an incorrect or non-existent record ID will result in an empty or error response. Double-check the record ID from the CRM URL or via a records API call before invoking the tool. [5]

  • Edition limitations. Zia's Prediction features are not available on all Zoho CRM plans. If the tool returns no data or an access error, verify that your organisation's edition includes Zia Prediction. [1][2]

  • Insufficient training data for churn scores. Zia's churn predictions rely on analysing multiple touchpoints. If a record is new or has limited interaction history, the prediction score may be low-confidence or absent. Enriching Zia with usage data from external tools (via API or file import into a custom module) can help address this. [3]

  • Forecast anomaly alerts vs. record predictions. Zia's anomaly detection in forecasts operates at the team/pipeline level and is surfaced as notifications to sales leaders — it is not the same as calling getziapredictions on an individual record. Mixing up these two features is a common source of confusion. [4]

What to check


  • Verify the tool response contains a valid prediction score for the module and record ID you supplied — a null or empty result usually points to an edition restriction or insufficient record data. [1][5]
  • Confirm usage data is mapped correctly if you are relying on churn predictions; check that your custom module events are recognised by Zia so scores reflect the full picture. [3]
  • Review Zia's Prediction section in the Zoho CRM Knowledge Base to ensure any new capability updates (such as expanded churn scoring signals) are accounted for in your integration. [1][2]

---


*Beam Help is an independent expert support resource for Zoho products and is not official Zoho support. Always validate against your specific Zoho CRM edition and the latest Zoho documentation.*

Sources cited

  1. [1] CRM-V2 | Zia | Knowledge Base
  2. [2] CRM | Zia | Knowledge Base
  3. [3] Zoho CRM Q2 2025 Update
  4. [4] Q1 2025 Update
  5. [5] get_zia_predictions
  6. [6] How do we change system field names?
  7. [7] Remove "Subject" as a required field on Quotes
  8. [8] Consolidate your workflows and be more productive - Introducing Multiple Conditions in Workflow Rules
Retrieve ZIA Predictions | Beam Help